Choi Eun-soo, CEO of IntelliVIX: “Trustworthy AI Protects People and Property”

AI Records 24-Hour CCTV Surveillance
Korea’s First ISO/IEC 42001 Certification Secured

“Awareness of Major Disaster Prevention Must Be Strengthened”

“What happens if a theft occurs and the system fails to detect it? What if a fire breaks out and the alarm does not go off? Trust is immediately lost. We must build a safe society based on reliable AI and accurate detection performance.”

Choi Eun-soo, CEO of IntelliVIX, a video analysis AI company, emphasized the importance of AI reliability while explaining why trust in artificial intelligence is critical. His strong focus on reliability stems from the fact that AI is directly used to protect people’s lives and property.

This commitment has been validated through internationally recognized certifications. Last year, IntelliVIX became the first company in Korea to obtain ISO/IEC 42001, a global standard that evaluates accountability, reliability, and ethics in artificial intelligence. Following this achievement, the company won the Good AI Awards 2024, which recognize trustworthy and high-quality AI technologies in Korea, marking its third consecutive win.

Trust is not built overnight. IntelliVIX has focused on AI research and development for 25 years. Approximately 70% of its workforce consists of technical personnel, and the company holds 35 patents related to video analysis alone. Its video detection accuracy reaches 99.55%. In the previous year, IntelliVIX invested ₩12.2 billion, accounting for 35.8% of its revenue, in research and development. CEO Choi stressed that continuous technological advancement is essential to building reliable AI.

◇ Public, Industrial, and Defense Sectors: Expanding AI CCTV Applications

IntelliVIX’s AI CCTV solutions are actively deployed across public safety, industrial safety, and national defense sectors. Currently, approximately 10,000 AI CCTV units are operating in local governments nationwide, including Gwangjin-gu in Seoul, Hwaseong, Gimcheon, Goyang, and Bucheon. “We protect public safety around the clock by monitoring fires, theft, falls, traffic accidents, wrong-way driving, and illegal waste dumping,” Choi explained. “AI monitors 10,000 cameras simultaneously, eliminating blind spots in control.”

Demand is also rapidly increasing in industrial sites. Following the enforcement of the Serious Accidents Punishment Act, companies have strengthened their safety management obligations. Major manufacturing and construction firms, including POSCO, LG Electronics, Samsung C&T, Samsung Engineering, LG Chem, Hanwha Chemical, and GS Caltex, have introduced IntelliVIX’s industrial safety AI. The system detects falls, entrapment, fires, missing protective equipment, and gas leaks in real time, reports incidents to companies, and automatically alerts emergency services.

The company is also expanding into the defense sector. As South Korea’s troop levels have decreased from 600,000 to 400,000, surveillance gaps along GOP barbed-wire fences and coastal areas have widened. “During snowfall, fog, or nighttime conditions, it is difficult for guards to detect intrusions,” Choi noted. “AI guards can monitor these areas 24 hours a day and deliver detailed information in real time.”

The VIXallcam is currently undergoing field testing at GOP barbed-wire fences and coastal posts. It is an AI camera capable of accurately detecting objects at distances of up to 200 meters, even under harsh conditions such as snow, rain, fog, and darkness. In the event of an intrusion at night, it can transmit detailed information in real time, such as the number of individuals, movement direction, speed, and posture. VIXallcam was selected as a CES 2026 Innovation Award honoree, gaining recognition for its global technological competitiveness. IntelliVIX is also developing a 500-meter detection technology. “Once the 500-meter-class technology is realized, applications will expand further to areas such as maritime collision prevention and port monitoring,” Choi said, adding that it will contribute to innovation in the defense sector.

Overseas expansion is also progressing. IntelliVIX has already supplied AI CCTV solutions to Japan, the Philippines, and Taiwan, and is currently collaborating with partners in Indonesia.

◇ From Video to Text: Advancing AI Capabilities

Recently, IntelliVIX developed a system that enables AI to describe CCTV footage in text form. The core technology behind this capability is Gen AMS. The platform also supports natural language search, made possible by converting video data into text. Gen AMS is equipped with a proprietary Vision Language Model (VLM) that combines vision AI with large language models.

“In the past, control personnel had to manually record incidents after detecting abnormalities on CCTV screens,” Choi explained. “Now, AI analyzes the video, generates textual descriptions of events, and automatically prepares reports.”

IntelliVIX’s VLM has been trained on 13,810 hours of video data and 50 million images. It supports detection of 91 object categories, encompassing approximately 500 million objects and 200 million data augmentation samples. “Extensive data training is the key to reducing false alarms and improving accuracy,” Choi added.

The company is also expanding into the field of physical AI. A four-legged patrol robot equipped with AI is being developed to monitor hazardous areas around the clock. “We are developing what we believe will be the world’s first AI patrol robot capable of generating reports,” Choi said. “When incidents such as fires, thefts, or collapses occur, the robot autonomously moves to the site, records the situation, and prepares a report.” A prototype is scheduled for release in March next year.

To support these developments, IntelliVIX is collaborating with NVIDIA to enhance its Vision Safety System (VSS). “We will accelerate performance using GPUs while also developing lightweight models compatible with domestic NPUs to enable broader adoption,” Choi said. “Only when feedback is delivered within 0.23 seconds can immediate responses be ensured.”

◇ “Preventive Awareness and Investment Must Increase”

Throughout the hour-long interview, Choi repeatedly emphasized the importance of “trust,” reiterating that “AI is not merely a technology, but a technology that saves lives.”

Large-scale incidents such as the Itaewon crowd disaster and the Arisel fire could have been prevented. “These were tragic accidents that might have been avoided if AI CCTV systems had been installed in advance,” he said. “Although awareness has improved since the enforcement of the Serious Accidents Punishment Act, many industrial sites are still responding reactively rather than proactively.”

https://digitalchosun.dizzo.com/site/data/html_dir/2025/11/24/2025112480089.html

​